Contents
- 1. The F&I Procedure Published by COPE
- 2. Minds and Machines: The Complaint That Had Received No Answer
- 3. April 10, 2026: A New Explanation from the Editor-in-Chief
- 4. April 20, 2026: The COPE F&I Response
- 5. Discover Artificial Intelligence: The Original Review and the Editor's Explanation
- 6. F&I Non-Intervention Based on ‘Editorial Independence’
- 7. Additional Verification Previously Conducted by the F&I Itself
- 8. 2026: Refusal to Escalate the Case to the Trustee Board
- 9. Summary of the Objective Facts
- 10. Discussion: What Does COPE Mean by ‘Editorial Independence’?
- 11. Noncommercial Advice from UTIE Instruments Inc. to COPE
1. The F&I Procedure Published by COPE
The COPE Regulations define the F&I Subcommittee's responsibility as reviewing and facilitating the resolution of concerns about the ethical practices of COPE member journals or publishers. Under COPE's published F&I flow, receipt of a procedural summary or response from a member journal or publisher is not itself the end of the process.
After that, the F&I Subcommittee is to:
- ask questions of, or provide comments or recommendations to, the complainant and the COPE member;
- seek additional information from the member where necessary;
- have the responsible F&I member and the Facilitation and Integrity Officer review all correspondence;
- attempt to achieve understanding and resolution between the parties; and
- determine whether the process was adequate and whether corrective measures or improvements are required.
The published procedure further states that, if the complainant does not agree with the final outcome, the matter will be reviewed by the Trustee Board. If serious concerns are identified, the F&I may recommend sanctions to the Trustee Board.
2. Minds and Machines: The Complaint That Had Received No Answer
April 7, 2026: COPE's Initial Inquiry
Regarding the complaint about the handling of a submission to Minds and Machines, Iratxe Puebla sent the following email to the author.
Primary source: Iratxe Puebla to the author
Dear Dr Naito,
I am writing regarding your submission to COPE involving concerns about the handling of your manuscript submission to the journal Minds and Machines. In your submission you indicated that you had not received a response from the journal about the rejection letter that did not include reviewer comments. Can you please indicate whether you have now received any response from either the journal or Springer Nature?
Thank you for this clarification.
With best wishes,
Iratxe
Iratxe Puebla
Facilitation and Integrity Officer
Committee on Publication Ethics (COPE)
April 7, 2026: The Author's Response
The author explained that, despite receiving a notification indicating rejection after peer review, no reviewer comments existed and no response had been received from the editor, the editorial office, or the Springer Nature ethics team. The author also submitted to COPE the email received from the Assistant Editor at that time.
Primary source: Springer Nature Assistant Editor
Dear Dr. Naito,
I have contacted the editor in this regard and will get back to you with an update. Thank you very much.
Best regards,
Christiani
Springer Nature
3. April 10, 2026: A New Explanation from the Editor-in-Chief
The author then received a response from Mariarosaria Taddeo, Editor-in-Chief of Minds and Machines. Taddeo said that the notification had been an error: the manuscript had in fact been desk rejected before review, and she had merely selected the post-peer-review rejection email template by mistake.
Primary source: Mariarosaria Taddeo to the author
Dear Dr Naito,
I am sorry for the long time it took to assess your paper.
Unfortunately, we have had some serious problems with the executive editor who should have followed the review process of your paper.
There are no reviews attached to the email because the paper has been rejected before review. When sending the editorial decision, I mistakenly selected the wrong editorial decision (I selected ‘reject after reviews’, I should have selected ‘reject before reviews’), prompting a different email template that includes reviewers’ comments in the text.
I desk rejected your article; the procedure does not require any review or comments to be shared with the author.
I hope you can accept my apologies for the delay in the review process and for this mistake.
Best
Mt
Professor Mariarosaria Taddeo
Professor of Digital Ethics and Defence Technologies
Oxford Internet Institute, University of Oxford
On the same day, the author not only forwarded this response to COPE but also submitted records that contradicted the explanation.
Primary source: the author to Iratxe Puebla
Dear Iratxe,
I just received a direct reply from the Editor-in-Chief, stating that the manuscript was indeed rejected by the desk, but that they mistakenly sent the post-peer review email.
However, the attached screenshot of the Editorial Manager portal, saved on 20 January 2026, shows the status as "Reviewers Assigned" (29 Dec 2025). The same status was also confirmed on the Springer Nature author portal on the Jan 2026. In other words, five months after submission, the editor is retroactively claiming a desk rejection for a manuscript that had already entered the review process.
I attach her email and the screenshot. Thank you for your involvement in this matter.
Best regards,
Hiroki Naito
The information submitted to COPE did not merely establish that the Editor-in-Chief had finally provided some answer. At the same time, the author submitted specific rebuttal evidence: although the answer described a pre-review desk rejection, Editorial Manager saved on January 20, 2026 displayed ‘Reviewers Assigned’ dated December 29, 2025. The author therefore asked COPE to intervene.
4. April 20, 2026: The COPE F&I Response
Iratxe Puebla responded as follows.
Primary source: Iratxe Puebla to the author
Dear Dr Naito,
Thank you for sharing the information you received from the Editor-in-Chief of the journal Minds and Machines. The editor has now provided an explanation regarding the handling of your manuscript and noted that the rejection was handled by the editor without external reviews.
These clarifications address the items you submitted to COPE regarding the editorial process followed for the manuscript. As previously noted, the Facilitation and Integrity subcommittee cannot interfere with editorial decisions; the editorial decision for rejection, and whether this is reached with or without external reviews, is beyond the scope of the subcommittee and thus, the subcommittee cannot pursue a review of those items.
With best wishes,
Iratxe
Iratxe Puebla
Facilitation and Integrity Officer
Committee on Publication Ethics (COPE)
Here, the F&I stated, ‘The editor has now provided an explanation,’
and, ‘These clarifications address the items you submitted to COPE.’
5. Discover Artificial Intelligence: The Original Review and the Editor's Explanation
During the same period, a separate COPE F&I case concerned discrepancies between the primary record and the editor's account of peer review at Discover Artificial Intelligence.
In the decision email dated November 28, 2025, handling editor Da Tao notified the author as follows.
Primary source: editorial decision
“Both reviewers recommended rejection.”
However, the actual comment from Reviewer 2 reproduced in the decision email stated:
“This manuscript addresses an important emerging issue but requires major revisions to meet publication standards.”
and immediately afterward stated, ‘The primary concern is the apparent fictional nature of the described events.’
The detailed review report attached to the email further stated the following decision at the outset.
“Overall Recommendation Major Revisions Required”
The report also described the premise that GPT-5 did not exist as follows:
“The study's fictional premise (GPT-5 does not exist as of 2025)”
It stated that this substantially diminished the contribution of the research. It further said:
“Using GPT-5 (which doesn't exist) as a co-coder is impossible”
“The described “backlash” and hashtag campaigns are fictional”
“This needs immediate clarification: Is this a speculative/scenario analysis? If so,
it must be clearly labeled as such throughout”
The requested revisions were classified as a Fundamental Premise Problem and listed determinations that GPT-5 had not been released and that the events, methodology, and data were entirely fictional. The final determination in the detailed report was:
“Recommendation: Major revisions required before reconsideration”
This was the report's final recommendation.
Accordingly, the actual record showed that:
- Reviewer 2 did not recommend rejection, but instead selected Major Revisions Required.
- Reviewer 2 expressly identified the mistaken assertions that the article's subject and data were fictional and that its methodology was impossible as central issues, using the terms ‘The primary concern’ and ‘Fundamental Premise Problem.’
February 6, 2026: The Author Again Provides Reviewer 2's Actual Recommendation to COPE
Primary source: the author to Iratxe Puebla and Laura Whitehead
Dear Iratxe,
I would like to provide a brief but critical fact-check regarding Ms. Laura Whitehead's reply.
In her very first sentence, she states
"...after receiving two reviews who both recommended reject."
However, as I have previously shared with you (and reported to Springer since November), the actual record of Reviewer 2(AI) clearly states: "Recommendation: Major revisions required before reconsideration." It is deeply concerning that after more than two months of investigation, Springer Nature Research Integrity provides a report to COPE that contradicts the evidence. This is no longer about the handling editor’s "feeling." It is about a factual falsification of editorial records being presented as a formal investigation result.
Best regards,
Hiroki Naito
April 20, 2026: COPE Explicitly Recognizes the Question of Consistency Between the Rejection and the Reviewer Recommendations
Primary source: Iratxe Puebla to Laura Whitehead, with the author copied
Dear Laura,
Thank you for the information you previously provided about Hiroki Naito’s concerns in relation to the handling of their submission to the journal Discover Artificial Intelligence.
A member of the Facilitation and Integrity subcommittee has looked at the concerns and would appreciate additional clarification from you regarding the procedural aspects of the author’s appeal request and the handling of the concerns that one of the review reports may have been AI generated.
We note that Dr Naito has also raised concerns about the decision for rejection and how this reconciles with reviewer recommendations, but the subcommittee does not interfere with specific editorial decisions, and thus the subcommittee will not review or comment on those points.
Dr Naito noted that they received a letter from the journal stating that an appeal would only be considered if 'important additional data' were provided, and a letter from the Springer Nature Research Integrity Group indicated that the handling editor reviewed the appeal in this case, focusing on the suitability of the manuscript for publication. Can you please provide information on the following points related to the handling of the author's appeal/complaint:
Does the journal consider appeals based on peer-review integrity concerns or concerns about decisions, in the absence of new data?
Was the Editor-in-Chief involved in evaluating the appeal (in addition to the handling editor), as is described in the journal's policy?
Please comment on whether and how the journal or the Springer Nature Research Integrity Group followed up on the presenter's concerns about (i) AI use in peer review, (ii) the identity of Reviewer 2, and (iii) a potential breach of the unpublished submission's confidentiality.
The journal's policy on AI in peer review asks peer reviewers not to upload manuscript contents to generative AI tools but also indicates that any AI use must be disclosed in the peer-review report. Can you please clarify what types of AI use in peer review would be considered acceptable by the journal?
Many thanks for helping us address Dr Naito’s concerns. We look forward to hearing from you.
With best wishes,
Iratxe Puebla
Facilitation and Integrity Officer
Committee on Publication Ethics (COPE)
In this email, the F&I itself expressly recognized that a concern had been raised about how the ‘decision for rejection’ reconciled with the ‘reviewer recommendations.’ At the same time, it stated that it ‘will not review or comment’ on that point.
6. F&I Non-Intervention Based on ‘Editorial Independence’
In these cases, the F&I took the position that it could not intervene in editorial decisions themselves because of editorial independence.
Primary source: Iratxe Puebla to the author, January 21, 2026
Dear Dr Naito,
Thank you for sharing the communication you received from the journal Discover Artificial Intelligence. The editor has commented on aspects of the scholarly content of the manuscript and the grounds for the rejection, those aspects are beyond the scope of the Facilitation and Integrity subcommittee as COPE cannot comment on editorial decisions.
With regard to the concerns that one reviewer provided an AI-generated review, the journal has indicated they are following up on that so at this stage, COPE must await the outcome of the journal and Springer Nature’s follow up on that concern.
With best wishes,
Iratxe
Iratxe Puebla
Facilitation and Integrity Officer
Committee on Publication Ethics (COPE)
Springer Nature's own public explanation of editorial independence centers on keeping editorial decisions independent from commercial concerns and preventing commercial or executive functions from interfering with those decisions. COPE's own guidance likewise describes editorial independence as the principle that editorial decisions should be based on the quality of submissions and appropriate peer review, free from political, financial, or personal influence. Editorial independence therefore is not a principle that justifies declining to verify whether the contents of a peer-review report are consistent with objective facts.
By contrast, the disputed matters documented in this article included questions that can be established as objective facts, independently of an editor's value judgment:
- whether the reviewer recommended rejection;
- what the reviewer itself described as the primary concern;
- whether the subject and data of the article were fictional;
- whether the article's methodology was impossible; and
- whether the explanation that the manuscript was desk rejected conflicts with the system record stating ‘Reviewers Assigned.’
These were not merely matters of editorial judgment.
However, the two cases were not identical in nature. In the Discover Artificial Intelligence case, the F&I Subcommittee itself expanded the scope of editorial independence. In the Minds and Machines case, despite concrete rebuttal evidence in the form of the Editorial Manager record having already been submitted, Puebla treated the matter as addressed solely on the basis of the Editor-in-Chief's earlier explanation.
7. Additional Verification Previously Conducted by the F&I Itself
Puebla's own handling of an earlier complaint confirms that the F&I does not ordinarily close a case merely upon receiving some response from a member journal. In 2020, Marc Edwards raised concerns about an abstract by Maki et al. published in the International Journal of Infectious Diseases. Even after receiving the journal's response, Puebla considered the follow-up inadequate and required further action.
On August 26, 2020, Puebla notified the journal that, unless it took the necessary additional action:
“we will have to raise the case to the COPE Trustee Board”
Puebla gave this notice on August 26, 2020.
According to COPE's own final report, the F&I subsequently referred the case to the Trustee Board, and three Trustee Board members reviewed it. They concluded that no formal sanction was necessary, but additional feedback was provided to the journal.
Thus, in that matter:
- The member provided a response.
- The F&I reviewed the response.
- It found the response insufficient.
- It requested further action.
- It actually escalated the case to the Trustee Board.
8. 2026: Refusal to Escalate the Case to the Trustee Board
On August 6, 2026, however, when the Company requested higher-level review by Nancy Chescheir and the COPE Trustees in the Discover AI F&I case, the F&I responded as follows.
Primary source: F&I response dated August 6, 2026
“We decline your request to have this matter reviewed by Nancy Chescheir and the COPE Trustees: per COPE’s standard procedures we do not reconsider or escalate cases which the subcommittee has already reviewed.”
In other words, the F&I stated:
‘Because the F&I Subcommittee has already reviewed the case, it will not be escalated to the Trustee Board.’
This was the explanation given.
However, COPE's currently published F&I flow expressly states that, if the complainant does not agree with the F&I's final outcome:
“the concern/issue will be reviewed by the Trustee Board”
Moreover, in the 2020 matter described above, an F&I case handled by Puebla herself was in fact referred to the Trustee Board after F&I review and was reviewed by three Trustees.
This is the opposite of the explanation given in 2026.
9. Summary of the Objective Facts
The primary materials reproduced above establish at least the following facts. In the Minds and Machines case, the F&I first asked the author whether a response had been received from the journal or Springer Nature. The author subsequently received an answer from the Editor-in-Chief but reported, with specific records, that the answer conflicted with the actual Editorial Manager record. The F&I nevertheless treated the items submitted to COPE as already addressed because the Editor-in-Chief had provided an explanation, and ended its involvement in the concern.
In the Discover Artificial Intelligence case, the author submitted evidence concerning whether the reviewer had recommended rejection, what the reviewer itself described as the primary concern, and the objective facts that GPT-5 existed, that the article's subject and data were not fictional, and that its methodology was not impossible. COPE stated that it could not intervene in editorial decisions and would limit its involvement to procedure. On April 20, 2026, the F&I expressly recorded in its inquiry to Springer Nature that the author had raised the question of how the decision for rejection reconciled with the reviewer recommendations, but stated that ‘the subcommittee will not review or comment’ on that point.
In both cases, even after specific rebuttal evidence showed that an explanation directly contradicted the primary record, the F&I and Iratxe Puebla treated the mere existence of that explanation as sufficient and declined further involvement on the basis of editorial independence.
10. Discussion: What Does COPE Mean by ‘Editorial Independence’?
The original and accurate meaning of editorial independence is to guarantee that commercial or management functions do not improperly interfere with editorial decisions. It is an internal governance principle designed to separate authority between different parts of an organization.
It emphatically does not mean that, no matter what objectively false statements or direct contradictions of a reviewer's own record are contained in a decision by an editor or publisher, an external body such as COPE must never examine whether those statements are true.
On the contrary, Springer Nature's own Editors' Code defines the purpose of peer review as accurately assessing the validity of research and expressly requires the process to provide constructive and critical evaluation, including whether conclusions are properly supported by the results.
This cannot be explained simply by saying that Iratxe Puebla did not understand COPE's basic rules. Puebla is herself a researcher—although a 2003 paper on which she was first author was retracted in 2025 for image splicing—and has worked for many years in academic publishing and publication ethics. She now serves as COPE's Facilitation and Integrity Officer. In an earlier case, she personally notified the parties that an F&I matter would be escalated to the Trustee Board and actually advanced that procedure.
Puebla therefore took editorial independence, a principle that is limited in scope and serves a specific purpose, and expanded it in an exceptionally convenient and arbitrary manner, deliberately misusing it as though it were immunity that completely protects any objective falsehood by a publisher or editor from external verification. Her research paper was retracted, and as a publication-ethics officer she applies procedures that depart from the standards of COPE's own published process. The one thing that has remained consistent over the years is her failure to follow the rules.
If Iratxe Puebla or COPE applies a standard under which:
a response is accepted merely because it exists, even when it directly contradicts the primary record,
then that standard should not remain undisclosed as it is now. If this is not the standard, an explanation is required as to why COPE's procedures for additional verification, resolution, and Trustee Board review were not applied. Iratxe Puebla and COPE have so far remained silent on this question.
11. Noncommercial Advice from UTIE Instruments Inc. to COPE
From a ‘Self-Help Group’ to an External Authority
In 2014, former COPE Chair Elizabeth Wager described COPE's principal function as providing advice and guidance to editors and publishers. She called COPE a ‘self-help group’ for editors and expressly stated that COPE was neither a regulatory body nor an organization that itself investigated suspected misconduct. Today, however, COPE membership is widely used within the industry as an external indicator of the credibility of publishers and journals. Springer Nature likewise describes COPE membership as one indicator that distinguishes legitimate publications from poor-quality outlets. The COPE Regulations define publishers as fee-paying members, and Springer Nature is a fee-paying publisher member of COPE. Its Research Integrity Director, Chris Graf, was one of COPE's early Co-Chairs, and the company states that it continues to work closely with COPE.
In cases brought to COPE, one party is a dues-paying COPE member, and often the person handling the matter for that member organization has themselves been deeply involved in COPE's governance. Under this conflict-of-interest relationship, when a member publisher is the subject of a complaint, whether the officer accepts a false report from the member at face value or examines it against the factual record becomes a decision point that goes to the very purpose of COPE.
In the two cases examined in this article, the former course was chosen rather than the latter. In both the Minds and Machines case and the Discover Artificial Intelligence case, Iratxe Puebla treated the concerns submitted to COPE as resolved merely because an explanation existed, even though the member publisher's explanation directly contradicted the primary records submitted by the author. By contrast, in a separate 2020 case in which Springer Nature was not a party, Puebla herself sought further explanation after receiving a response from the member journal because she considered the response inadequate, and ultimately referred the case to the Trustee Board. COPE has not explained why the same officer applied such markedly different levels of scrutiny under the same procedure.
On the Springer Nature side, Chris Graf, the company's Research Integrity Director and a former COPE Co-Chair, used a COPE report containing the factually incorrect claim that an Editor-in-Chief had been involved as the basis for asserting that the peer-review process was appropriate, and refused the author's request for a further investigation. The details are examined in the separate article ‘Is a Fictitious Investigation by Springer Nature's Research Integrity Team a Problem?’
COPE's published F&I procedure is designed as a systematic framework that includes obtaining additional information, building understanding between the parties, considering corrective measures, and escalating matters to the Trustee Board. The fact that this procedure was not applied only in cases in which Springer Nature was a party must be understood as selective application of the procedure in a manner that benefits a commercial member organization with which COPE has a conflict-of-interest relationship.
‘Editorial independence’ is originally a principle intended to prevent commercial or management functions from interfering with editorial decisions. In this case, Iratxe Puebla repurposed the principle as a basis for shielding statements made by publishers or editors from factual verification by outsiders or by the author. Iratxe Puebla was also the first author of a paper involving research misconduct, but she now holds an important position reviewing publication ethics. How COPE corrects the actions she took as an individual officer in the two cases examined in this article, contrary even to COPE's own published procedures, will be an important test if COPE is to avoid being described as merely a tool for laundering misconduct by member publishers.